# matrix

This topic is 3459 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

## Recommended Posts

hy. I have a matrix in collada file like this: <matrix>1 0 0 0.47476 0 0 -1 -0.480625 0 1 0 -0.122112 0 0 0 1</matrix> how ransform it for apply the position transform in directx? THe matrix must be colum major or row major(is correct the term?) And how transform a 16 element array tro a D3DXMATRIX? THanks.

##### Share on other sites
D3D uses row vectors/matrices and row major order. Collada uses column matrices but in row major order (if I remember correctly). So you cannot simply load a Collada matrix into a D3D matrix memory and be happy.

E.g. a sequence of elements in Collada like
{ a b c d e f g h i j k l m n o p }
has an equivalent in D3D like
{ a e i m b f j n c g k o d h l p }
(if I made no error ;) )

The mathematical operation to correct this is the transposition. So the simplest way may be to load the D3D matrix storage with the Collada co-efficients and to apply D3DXMatrixTranspose(...). Of course, you can already use the above scheme and put the co-efficients one by one into the correct place during reading the Collada matrix.

##### Share on other sites
and , i have a z_up document, and i would transform to a y_up , then i must invert y with z and negate y:
x,z,-y how do the same with matrix?
thanks.

##### Share on other sites
To map (x,y,z,w) to (x,z,-y,w) you need a (row) matrix as follows:
[ 1  0  0  0 ][ 0  0 -1  0 ][ 0  1  0  0 ][ 0  0  0  1 ]

Notice please that I've completed the vectors by the homogeneous co-ordinate w, since D3D matrices are homogeneous (i.e. 4x4 elements). You should take the time and perform the multiplication on paper once to see how it works.

##### Share on other sites
thanks , but what i do with this matrix?

i have a direct3d matrix ?
dxTransp._11 = vPosition[posRot][0];dxTransp._12 = vPosition[posRot][4];dxTransp._13 = vPosition[posRot][8];dxTransp._14 = vPosition[posRot][12];dxTransp._21 = vPosition[posRot][1];dxTransp._22 = vPosition[posRot][5];dxTransp._23 = vPosition[posRot][9];dxTransp._24 = vPosition[posRot][13];dxTransp._31 = vPosition[posRot][2];dxTransp._32 = vPosition[posRot][6];dxTransp._33 = vPosition[posRot][10];dxTransp._34 = vPosition[posRot][14];dxTransp._41 = vPosition[posRot][3];dxTransp._42 = vPosition[posRot][11];dxTransp._43 = -vPosition[posRot][7];// for the translations i'm inverted they/xdxTransp._44 = vPosition[posRot][15];

for the translate all is ok( i flip 42 with 43 and negate 43), but for the scale and the rotate what change i do(what i flip and what negate) in the matrix?
Thanks

##### Share on other sites
Quote:
 Original post by haegarrTo map (x,y,z,w) to (x,z,-y,w) you need a (row) matrix as follows:[ 1 0 0 0 ][ 0 0 -1 0 ][ 0 1 0 0 ][ 0 0 0 1 ]Notice please that I've completed the vectors by the homogeneous co-ordinate w, since D3D matrices are homogeneous (i.e. 4x4 elements). You should take the time and perform the multiplication on paper once to see how it works.

I'm not understand , what I do with this matrix?I must multiply it for my matrix?
why?
Thanks.

nobody?

nobody?

nobody?

1. 1
2. 2
3. 3
frob
20
4. 4
JoeJ
20
5. 5

• 11
• 11
• 12
• 13
• 9
• ### Forum Statistics

• Total Topics
632213
• Total Posts
3004848

×